Microsoft redobla su inversión en IA

Microsoft intensifies its investment in AI as it aims to reach $175 billion in capital expenditures this year, a historic amount for the tech giant. The company continues to build data centers despite investor doubts about whether the massive investment will yield expected returns. Like other U.S. Big Tech companies, Wall Street is closely monitoring quarterly results to determine if Microsoft's AI monetization is progressing at a good pace.

Microsoft's bet on AI is part of an unprecedented competition among major tech firms. Amazon, Google, and Meta have also accelerated spending on AI infrastructure, as computing capacity has become a key competitive factor in the sector. The company has added 31 new data centers in the June quarter, bringing the total to 88 for the year, and added an additional gigawatt of power, all while aiming to double its total capacity within two years.

Microsoft acknowledges demand prediction difficulties but stresses that it currently exceeds available capacity, necessitating accelerated data center construction. Financial Director Amy Hood recently addressed questions from analysts about potential overinvestment in AI, stating that Microsoft has room to adjust the investment pace if demand changes.

Two-thirds of the capital expenditures are allocated to short-term assets, primarily chips. Microsoft also maintains that it can scale data center construction based on real market signals. The company holds a $678 billion pipeline of future contracted revenues, a significant portion tied to OpenAI, where Microsoft holds a 27% stake.

Microsoft reported $24 billion in revenue from agreements with OpenAI, the ChatGPT owner, during the most recent fiscal year, representing around 7% of its total revenue of $331 billion. Microsoft presents AI as the growth lever for the future, stating that the technology is already transforming consumer and organizational productivity.

CEO Satya Nadella envisions an AI-driven offering encompassing cloud infrastructure, productivity applications, and solutions for agent, security, and software development. Nadella emphasizes that enterprises should not only consume external AI models but also build their own learning systems using internal knowledge. Microsoft is betting on offering businesses access to a wide range of AI models.

The company boasts over 11,000 AI models in Azure, including those developed by OpenAI, Anthropic, Mistral, and xAI, in addition to its proprietary technologies. Nadella asserts that the number of enterprise customers using models from multiple providers has increased fivefold since the beginning of the year. Microsoft's cloud business Azure is showing strong growth driven by AI services demand.

In the last quarter, Azure grew 43% year-over-year, and the company anticipates a 45% increase in the current quarter. Nadella revealed during a recent analyst conference that Azure's revenue surpassed $100 billion in the last fiscal year. UBS notes that while this performance reflects strong demand for AI services, it also highlights the robustness of traditional cloud services demand.

Investors are closely watching Copilot's growth, considering it a crucial test of Microsoft's ability to monetize AI investments among enterprise clients. The company has disclosed that Microsoft 365 Copilot has surpassed 30 million paid licenses, ten million more than the previous quarter. Microsoft expects the number of organizations with more than 50,000 licenses to multiply by more than seven over the next year, with references including AstraZeneca, Boeing, Infosys, Koch, and Procter and Gamble.

UBS projects that the increasing adoption of Copilot and price hikes in Microsoft 365 will accelerate the division's growth in the next fiscal year, further monetizing Microsoft's AI bet. Microsoft has reinforced its infrastructure with 31 new data centers in the latest quarter, adding 1 gigawatt of additional power. Its future contracted revenue pipeline reaches $678 billion, with a significant portion linked to OpenAI. Nadella argues that businesses should go beyond consuming AI models and build systems.